A cohort study is a longitudinal study that assesses correlations within a group of similar individuals. That is to say a study design, whereby a group of individuals are observed during a period of time and exposure to a factor is assessed in relation to a certain outcome. Observational cohort studies are commonly used in nutritional studies when the researchers are interested in how exposure to a particular dietary constituent or feeding behaviour alters the outcome of a disease or condition. For example, researchers might be interested in how flavonoids affect cardiovascular disease. In this case a group of individuals may be followed over time and after a set period of time assessed for the out come, in this case development of cardiovascular disease. The flavonoid intake of the subjects would then be analysed to see if exposure to flavonoids decreased the risk. Confounding variables are a common problem in cohort studies.
Cohort studies can differ slightly in their design, but the main divisions within the category are prospective and retrospective studies. In the case of prospective studies, a group of individuals without a disease or condition are observed for a period of time and medical records used at the end of the study to assess the outcome. As they are observed, data is collected and the associations between the variables of interest and the outcome are calculated using a correlation coefficient (R2) to assess how closely a regression line fits the set of data. A retrospective study differs from this in so far as the outcome is assessed in the present, and the study looks back over a previous time period to assess any association. The reason for performing a retrospective study is usually based on the availability of data from large nutritional surveys, or the re-evaluation of previous data based on new understandings or theories.